Wireless networks have recently received much attention. Support for mobility in Internet access is gaining significant interest as wireless/mobile communications and networking are proliferating, especially boosted by the widespread use of laptops and handheld devices.
Wireless LANs can be classified into two broad categories — Infrastructured and Infrastructureless (ad hoc). In certain wireless networks, one hop is needed to reach the mobile terminal. Ad hoc networks normally require a multi-hop wireless path from the source to the destination. Interest in such networks is increasing day by day.
Wireless networks can be broadly classified into two types:
- Fixed wireless networks.
- Mobile wireless networks.
Fixed wireless networks are mostly point-to-point and they do not support mobility (e.g., microwave networks, geostationary satellite networks). Mobile wireless networks , on the other hand, support mobility. They can be further classified into two types — infrastructured (cellular) and infrastructureless (ad hoc).
This classification is not very strict. A mobile ad hoc network can also be infrastructured: several mobile nodes can form an ad hoc network and can also be connected to a wired gateway for Internet access.
Due to the increase in both computing power of the mobile device and wireless link bandwidth, a user can use his or her mobile device to perform certain tasks that were earlier not possible.
In multi-hop mobile ad hoc networks, the focus will be on different type of routing protocols and medium access control techniques. Several issues need to be taken into account while considering a wireless network.
Quality of service (QoS) needs to be provided. If the delay produced in packet-switched networks can be removed to match them with the QoS of circuit-switched networks, then packet-switched networks would definitely be the choice for customers. This is not a simple problem.
It actually means integrating delay-sensitive applications such as voice, audio, and video along with delay-insensitive applications such as e-mail, fax, file transfer, etc. The network must be able to differentiate between these two kinds of traffic and process both of them efficiently and on time.
When the packet-switched network is a wireless LAN, then the situation becomes even more challenging. The TCP/IP stack was not designed with mobility in mind. Some solutions to this problem have been proposed, such as Mobile IP, etc. On the wireless network side, the key characteristics are the following:
- Mobility of the users.
- Bit errors in the wireless channels.
- Scarce wireless resources.
On the IP network side, there are two key problems:
- Lack of QoS support.
- Lack of data synchronization.
An important trend is the ongoing work and trials on a mobile device with multiple wireless communications interfaces.
The rationale behind these efforts is that although there are diverse wireless/mobile communications technologies with their own characteristics (e.g., bandwidth and coverage), no single wireless/mobile communications standard is likely to be the norm in providing access to the Internet.
Rather, these technologies have more or less complementary features. For example, the advantage of cellular technology is global coverage, whereas its weakness lies in the bandwidth capacity (currently tens of kilobits per second for the data traffic) and the operational cost.
In contrast to this, IEEE 802.11 wireless LAN technology has a bandwidth capacity on the order of megabits per second with little operational cost but has a relatively short range of coverage. With this perception, there have been approaches to equip a mobile node (MN) with multiple wireless communications interfaces.
The most promising scenario so far is to equip the MN with a cellular network interface (e.g., GPRS, CDMA) and a wireless LAN interface (e.g., IEEE 802.11, HIPERLAN).
With this configuration, the MN can connect to the Internet through the cellular network interface when outdoors and also through the 802.11 interface when indoors, if there is an available 802.11 wireless LAN access point.
In this era of wireless technologies convergence, we should note that diverse wireless coverage of technologies (e.g., cellular, wireless LAN, and Bluetooth) overlap each other, and there are frequent vertical handoffs between different kinds of wireless networks.
To support seamless mobility to an MN with ongoing Internet connections, the most critical issue is handoff. There are mainly three kinds of handoffs, depending on the situation:
- Intradomain (subnet-level)
Here intradomain handoff and interdomain handoff are also called microlevel mobility and macrolevel mobility, respectively. The mobility management framework should deal with handoff in all of these cases, seeking to minimize disruption.
Especially, the interdomain handoff is likely to occur frequently in wireless technology convergence, and it will accompany considerable signaling traffic load and delay with the current solutions, to be detailed later on.
Although there is a consensus that Mobile IP will be used to provide macrolevel mobility management in wireless/mobile Internet access, there have been a number of proposals for the microlevel mobility issue.
Here, micro mobility is the case in which the MN is moving and thereby changing the point of attachment between subnets in the same wireless network (the same administrative domain).
Need For IP Addresses
Mobile networks have seen a tremendous growth in the last decade of the 20th century and the beginning of the 21st. Various mobile devices such as personal digital assistants (PDAs), wireless laptops, and cell phones have sprung up.
Users of these devices want the same services that are available to their wired counterparts (e.g., WWW service, e-mail service, file-sharing service, etc.). These services are provided through the widely used TCP/IP stack.
To provide the same services to the mobile devices, it is necessary that individual IP addresses be assigned to each such mobile device. This is a very challenging task due to the dynamic nature of such mobile networks, and their full integration with the Internet will take some time.
Packet switching is needed to perform this task as circuitswitched technology is too expensive for bursty data traffic typically present on wireless LANs. Using another addressing scheme than IP will make these networks incompatible with the Internet.
Also thousands of applications using the TCP/IP stack will be unusable. Therefore, to interconnect these networks with the Internet, IP addresses the need to be assigned to the nodes of the mobile wireless network. Most of the users now are mobile users as well as Internet users. They demand Internet access while they are on the move.
The original Internet normally used one traffic type for all applications, also known as best-effort traffic . However, for the integration of mobile networks with the mostly wired Internet, using one traffic type was not feasible. Moving MNs have different requirements from static nodes.
Different schemes may be employed to support QoS in mobile networks. QoS support is especially important in wireless IP networks, where resources are scarce and should not be wasted.
However, multiple traffic classes via the Type of Service (TOS) field in the IPv4 header format and via the Differentiated Services (DS) field in the IPv6 header format are supported.
In a wireless IP network, there would simultaneously exist different traffic types such as voice, audio, video, multimedia, and data. The applications may be classified into real-time (voice service) and non-realtime (e.g., e-mail and Web browsing).
Therefore, different traffic types have different QoS demands that should be satisfied accordingly. Different nodes have links with different bandwidths ranging from kilobits per second to gigabits per second.
Similarly, different applications are also heterogeneous. Some are real-time (e.g., VoIP, audio and video streaming), whereas some are not (e.g., Web browsing and e-mail).
A routing algorithm should have the following characteristics:
- Low computational complexity
Routing can be a challenging task in mobile wireless networks. The conditions of such a network are as follows:
- Fast-changing network topology.
- A destination wireless node that may be multiple hops away from the source wireless node.
- Nodes that may switch off/on, move in/out of the network anytime.
- The presence of sleeping nodes that may receive traffic just for themselves and not forward traffic to others.
These conditions do not allow the traditional dynamic algorithms implemented on wired networks to be implemented on wireless networks.
There are four mobility classes that support IP mobility:
- Pico mobility is the movement of an MN within the same base station (BS). The area extends to about 10 m in all directions, enveloping the user.
- Micro mobility is the movement of an MN across different BS with fast speed. This is supported by using link layer support.
- Macro mobility is the movement of an MN across different subnets within a single domain or region. This is typically handled by Mobile IP.
- Global mobility is the movement of an MN among different administrative domains or geographical regions. This is also handled by techniques such as Mobile IP (layer 3).
The goal is to provide uninterruptible connectivity during micro and macro mobility.
Frameworks For Supporting Mobility
There are several frameworks for implementing mobility in mobile wireless networks. The IETF has standardized two of them, Mobile IP and Session Initiation Protocol (SIP):
- Mobile IP provides application-layer-transparent IP mobility. All the traditional applications using TCP/IP stack can continue to function in a mobile wireless network. The drawback of Mobile IP is that it does not provide additional features such as authentication and billing.
- SIP is an application layer protocol that can establish, modify, and terminate multimedia sessions or calls. The drawback of SIP is that it does not support TCP connections and is also not a solution for micro or macro mobility. These drawbacks disallow the use of SIP as the protocol for supporting mobility.
Mobility in wireless networks is of two types — user mobility and terminal mobility. User mobility refers to the ability of end users to access calls and services from any terminal or any location. Terminal mobility may be referred to as the ability of a mobile terminal to access services from any location.
A user may carry the terminal to any location and access the services provided by the wireless network. It is estimated that the number of mobile subscribers will reach about 60 million by 2005.
Cellular networks are of many types: GSM, ADC, PDC, CT2, DECT, PACS, etc. 6,13 Location tracking or mobility management is the set of mechanisms by which location information is updated in response to endpoint mobility.
The address and identifier of the endpoint need to be maintained for this purpose. In most cellular networks, precomputed routes are used, whereas in infrastructureless mobile ad hoc networks, packet switching is used.
Every node computes its own routing table. In cellular wireless networks, usually central entities perform the function of coordination and control, whereas in mobile ad hoc networks, no such central entity exists. Related work on IP-level mobility management can be broadly classified into three categories:
- Work in the first category is focused on microlevel mobility (in short, micro mobility), mostly in the cellular network domain. As it takes considerable time to exchange the registration message between the foreign agent (FA) and the home agent (HA), most proposals in the first category have considered a special agent node in each administrative domain, which accommodates the local handoff within the administrative domain without contacting the HA of the MN.
- Work in the second category mainly seeks to reduce disruption and packet loss in handoff. The most-proposed schemes suggest a cooperating scenario between the old FA and the new FA.
- The third category is concerned with the authentication, authorization, and accounting (AAA) issue in regard to mobility in Internet service.
Modification of IP addresses for malicious purposes is very easy in a wireless ad hoc network, and this has to be taken care of. The IP security architecture is standardized by IETF and is also known as IPSec. With IPSec, security is directly applied to IP packets.
The concept of security association (SA) is used. SA consists of three parameters: destination address, cryptography protocol, and an identifier that separates multiple SAs with the same destination host but using the same cryptography protocol. SAs are unidirectional.
Two mechanisms are used to protect IP packets:
- The authentication header (AH) is used to protect the authenticity and integrity of an IP packet with a keyed cryptographic hash value.
- The encapsulating security payload (ESP) that transports encrypted IP packets, ensuring confidentiality and authenticity (optional), as well as integrity of the packet payload.
Routing Protocols For Implementing Mobility IP
Many routing protocols, i.e., Ad hoc On-Demand Distance Vector (AODV), Temporally-Ordered Routing Algorithm (TORA), Dynamic Source Routing Protocol (DSR), Destination-Sequenced Distance-Vector (DSDV) are proposed in the literature to provide routing services for the wireless network.
These protocols act as providing the necessary service for the wireless network. DSDV can be implemented on a wireless LAN to provide IP mobility. DSDV is a distance vector–routing protocol. DSDV allows for freedom-loop guarantee. But it also has periodic broadcast overhead because it is proactive.
DSR is another Mobile Ad-hoc Networks (MANET) routing protocol that has no periodic broadcast overhead, but the packet size is larger due to piggybacking of route information. AODV shares the advantages of DSR and distance vector–routing protocol.
Route discovery in AODV is similar to that of DSR. AODV maintains a routing table that contains the destination IP, destination sequence number, hop count, next hop, and lifetime.
The route information in the routing table is invalid if a RERR packet is received or the route lifetime is expired. Although TORA minimizes the reaction due to changes in network topology, the overhead is large due to IMEP (Internet MANET Encapsulation Protocol).
All IP-Mobile Network
Once it is possible to assign IP addresses to mobile devices as well as fixed devices, we would have an all-IP network. All the Internet services such as WWW, FTP, TCP, VoIP, etc., would then be available to all these devices.
The main advantage of such an all-IP network is efficiency (due to statistical multiplexing and the use of packet switching instead of circuit switching), low operational costs, and the transparency of IP technology to different kinds of services.
One service will cater to the demands of different types of users, e.g., a stock exchange update applet will show stock updates on personal computers, laptops, PDAs, and cell phones equally effectively. This will result in some very exciting possibilities.
TCP Over IP
TCP is the dominant standard for reliable data delivery on the Internet. Services such as WWW, FTP, and TCP traffic cover almost 95 percent of all bytes, 85 to 95 percent of all packets, and 75 to 85 percent of all flows. For real-time communication, mostly UDP is used.
TCP was originally created for wired packet networks; hence TCP does not show good performance over wireless networks. Congestion is the norm rather than the exception on wireless networks. Multiple packet losses inside a TCP segment means that TCP Selective Acknowledgment (SACK) may be appropriate for wireless networks.
There are profound advantages of IP mobility in wireless networks. We will consider two scenarios here:
- Consider the scenario of a massive battlefield. An enormous wireless network is deployed with thousands of soldiers, hundreds of tanks, and other vehicles acting as network nodes (routers).
Complete information regarding the battlefield is constantly transmitted to the headquarters, where it is analyzed and processed. Immediately, new instructions are sent back to the battleground, where they are received by the MNs, which adjust themselves accordingly.
- Another scenario is that of an earthquake disaster area. With widespread destruction, wired access to the earthquake area may not be possible. In such a situation, a wireless ad hoc network can quickly be deployed.
Other applications include man-made disasters, relief operations, military applications, car-based networks, sensor networks, the provision of wireless connectivity in remote areas, collaborative computing, and video conferences.
Future-generation mobile networks are going to be all-IP networks; thus, all types of information will be carried using IP packets. Although the basic infrastructure is in place, extensive research in the field of IP mobility and location management will be needed for full-scale deployment. Issues still outstanding are mobility management, security concerns, efficiency, etc.